The human MDR1 gene encodes a multispecific drug transporter, P- glycoprotein (Pgp), that prevents drug accumulation in resistant cells. Overexpression of MDR1 is sufficient for conferring multidrug resistance on otherwise normal cells. This suggests that MDR1 might be used in gene therapy to protect hematopoietic cells against chemotherapy-related myelotoxicity. Of 164 cancer- related gene therapy trials currently in force, nine incorporate the concept of hematopoietic cell chemoprotection; six of these use the MDR1 gene. Another application of MDR1 is to use it as an in vivo selectable marker to enhance the expression of linked foreign genes in transfected or virally transduced cells. Mouse experiments indicate that MDR1 can be chemoprotective and selectable in vivo, but attempts to use MDR1 as a chemoprotective agent in human gene therapy trials have, so far, been disappointing. Evidence will be provided suggesting that the reason for its poor in vivo performance in humans is that MDR1 is a stringent selectable marker that requires very high levels of P-glycoprotein, the MDR1 gene product, to mediate survival of transduced cells. Indeed, the most significant barriers to successful gene therapy with MDR1 appear to be transduction efficiency and gene expression--i.e., the number of cells that can be transduced and that can express high enough levels of MDR1 to survive selection. We hypothesize that MDR1 will serve as an effective in vivo selectable marker or chemoprotective gene only if the problem of stringency can be overcome. Four specific aims are designed to test this hypothesis and to develop the optimal strategy for overcoming selection stringency: 1) Determine if MDR1 selection stringency can be overcome by maximizing gene transduction efficiency and gene expression levels with state-of- the-art gene therapy tools. 2) Determine if selection stringency can be overcome by using mutant versions of MDR1/Pgp as the selectable marker. 3) Determine if selection stringency can be overcome by using a two-step selection strategy. 4) Determine if MDR1 can confer an in vivo survival advantage on human hematopoietic cells. Building on results in Specific Aims 1-3, this aim will use primary human hematopoietic progenitors to study the survival of MDR1-transduced cells in NOD/SCID and SCID- hu mouse model systems.